It's something of a moot point when I'm really arguing NASA wouldn't exist to pay for launches without tax dollars, but the non-binding launch contracts were structured like a research subsidy with many of the payments being associated with hitting r&d milestones rather than with an obligation to actually render services, and encouragement to go and find other customers. Nothing unusual about that: it's the way the space industry works.
That's not a subsidy. That's progress payments. Many industries other than space have progress payments.
In fact you see some private contracts which are exactly like that NASA contract: "We want to pay you to create a much cheaper product than we can already buy, and here are some early progress payments so you don't have to finance it yourself."
You may think this is a moot point, that's totally fine. The difference between subsidies or not is important to others.
You'll find private sector contracts which pay in instalments for hitting milestones, but not ones which give companies hundreds of millions of non-reimbursable research dollars without the obligation to buy or sell any actual services or transfer IP or equity. Phase I of COTS was a series of grants for hitting planning milestones with a right to terminate before Phase II. The procurement contracts were separate and subsequent to this.